“It’s troubling that nobody deigned to mention this was coming.”
At a public hearing at the end of November, Boston police and officials heard from activists and lawmakers concerned about transparency regarding surveillance technology in the city. The gunshot detection system ShotSpotter received particular criticism over accuracy and deployment, but officials defended the tech and said it was needed to increase public safety.
What city representatives did not say was that they were already planning on spending more money on gunshot detection, through a bid that was released two weeks after the hearing that also calls for retaining “all acoustic data” picked up by the contractor’s microphones.
Now, privacy experts and city councilors are saying they need more information about ShotSpotter—which has come under increasing scrutiny nationwide—and that law enforcement hasn’t met the standards of a new ordinance mandating more public reporting about surveillance.
“It’s very concerning and shocking that this hasn’t received much scrutiny,” said Fatema Ahmad of the Muslim Justice League of the recently-released bid. “The details are buried in these documents but rarely shared out loud.”
“It’s very troubling that people … at that hearing knew that this bid was being prepared and said nothing about it in a hearing designed to provide transparency and public accountability,” said Kade Crockford, director of the Technology for Liberty Project at ACLU Massachusetts. “It’s troubling that nobody deigned to mention this was coming. It’s not clear to me it’s a direct violation of the letter of the ordinance, but it certainly violates its spirit.”
City Councilor Kendra Lara, who is vice-chair of the Public Safety Committee that held the hearing, said councilors are still looking for information about who can access ShotSpotter and other surveillance equipment, along with specific locations where equipment is used in the city. Councilors need that data on surveillance before approving any reporting from police and other departments, she said.
“We’re not ready to move forward at this point,” Lara said. “We need to slow down and get this done properly.”
Under a city ordinance passed last year, police and other departments have to report what kind of surveillance they use and what policies govern that use to the City Council for approval. Any new surveillance technology must also be approved by the Council as well. While the Committee on Public Safety held that public hearing in November, it so far has not made any recommendations for approval or for not approving any surveillance practices in Boston, but could do so at the final City Council meeting this Wednesday.
The ordinance came after years of criticism of not just surveillance practices themselves, but lack of transparency around them. Boston and eight other cities initially planned to create a linked network of more than 1,000 surveillance cameras, but after the Boston Institute for Nonprofit Journalism revealed the proposal in 2021, then-Mayor Kim Janey pulled back on those plans. Later that year, WBUR reported that BPD purchased a cell-phone tracker using civil forfeiture funds, flying under the radar of any oversight.
And it’s not just BPD involved in watching Bostonians. Boston’s surveillance ordinance covers other departments, including the Office of Emergency Management—the city department that works as the agent handling grants for the Metro Boston Homeland Security Region that covers Brookline, Cambridge, Chelsea, Everett, Quincy, Revere, Somerville, and Winthrop, as well as Boston. Under the new ordinance, OEM has to submit its surveillance plans to city officials too.
OEM and MBHSR were behind the linked camera plan, and the agencies use federal Urban Area Security Initiative grants for their projects. Those grants funnel millions of dollars to law enforcement agencies, including to BPD to use on CIA-linked intelligence contractors. And according to a bid solicitation released on Dec. 5, OEM is preparing to deploy those funds for gunshot detection across the homeland security region.
[Additional Reading: With One Surveillance Plan Derailed, BPD Still Seeks To Add Cameras]
The bid does not name a specific gunshot detection system, but Boston departments have been using ShotSpotter for years. According to city financial data, BPD and OEM have paid ShotSpotter about half-a-million dollars a year over the last three years. BPD’s ShotSpotter use is specifically for Boston, but OEM’s use covers numerous areas. According to a 2020 document laying out USAI grant funding, MBHSR has managed gunshot detection systems in Boston, Cambridge, Chelsea, Everett, Revere, and Somerville for at least a decade.
Mayor Michelle Wu’s office did not respond to multiple requests for clarification and comment regarding the bid.
The current bid solicitation calls for a contractor to manage “project oversight, technical leadership, equipment acquisition, software installation, state server connectivity and maintenance” of a gunshot detection system for up to three years, covering eight square miles across six of the MBHSR’s locations. It also requires the contractor to detect gunshots across jurisdictional borders and send all alerts—“including fireworks”—to all jurisdictions as well. And it requires the contractor to hold onto all acoustic data, “not only limited to data that was classified as a gunshot,” for 30 days.
Ahmad said the OEM bid was another instance of a Boston agency not only increasing surveillance in the Hub, but in other cities whose residents are likely unaware of what’s going on.
“If Boston pays for other cities’ equipment, it is a concern. … it’s really a frustrating lack of transparency and public process,” Ahmad said. “Even if those cities don’t have surveillance ordinances, the reason those ordinances are popping up is we know departments are buying things secretly—BPD has a number of times. So to have another city and another city’s department doing this without going through processes or giving residents a chance is frustrating.”
“The fact that they had the bid going out … and didn’t think to mention it when ShotSpotter got all that attention [at the November hearing] was mind-boggling but not surprising,” Ahmad added. “If it’s not in direct breach of the ordinance, it’s very bad practice. … the timing is really wild.”
At the November hearing, privacy advocates raised specific concerns about ShotSpotter’s accuracy and its placement in minority communities. The company manages and monitors microphones that detect the sound of gunshots—supposedly weeding out similar noises—and sends reports to police departments that contract with them within a minute of the gunshot.
BPD Superintendent Marcus Eddings said a report by the firm Edgeworth Analytics backed up ShotSpotter’s claim of a 97.5% success rate in accurately identifying gunfire, and that the system helped police reach gunshot victims faster and increased community safety. And while BPD’s shot-spotting microphones are mostly in the Roxbury and Dorchester area, Eddings said, “If we had the budget we’d have ShotSpotter everywhere in the city.”
In a statement, a spokesperson for the company said, “ShotSpotter coverage areas are determined by police using objective, historical data on shootings and homicides to identify areas most impacted by gun violence. All residents who live in communities experiencing persistent gunfire deserve a rapid police response, which gunshot detection enables regardless of race or geographic location.” Adding, “We are trusted by police departments in over 135 cities nationwide and have a 99% customer retention rate, indicating that our system works well.”
Another review of ShotSpotter’s accuracy and Edgeworth’s analysis pointed out that Edgeworth determined accuracy largely through customer satisfaction—reports from police clients using the system. Alex Marthews, the founder of privacy advocacy group Digital Fourth, said a review of the Cambridge Police Department’s ShotSpotter usage showed that two-thirds of the time, officers were not able to verify a ShotSpotter report had actually heard gunfire. That could be because a shooting did take place and police could not find physical evidence, but also because the noise was inaccurately identified in the first place, Marthews said.
According to the ShotSpotter spokesperson, the company’s tech records sounds after three or more sensors detect a gunshot-like sound at the same time. The recording includes one second of ambient noise before and after the sound and audio is deleted every 30 hours; the spokesperson said a New York University Policing Project audit determined the risk of voice surveillance is “extremely low.” (According to its website, the New York University Policing Project’s partners include the NYPD, the LAPD, and the Cato Institute, and founders include the Charles Koch Foundation.)
However, OEM’s bid requires that “all data collected by the contractor and their equipment, including audio files, shall be retained and available to OEM and MBHSR jurisdictions upon request for retroactive analysis for a period of not less than thirty (30) days. This includes all acoustic data, not only limited to data that was classified as a gunshot.” Ahmad said she was concerned by that retention.
“This requirement, combined with the requirement that this integrates with the CIMS camera network, indicates that OEM is running a massive audio and visual surveillance network across these cities,” Ahmad said, adding the purpose of the ordinance and hearings like the one in November is to bring details like this to light. “That’s the reason that people pushed for the surveillance ordinance, so there is at least some public transparency.”
And according to specifications in OEM’s bid, the agency is looking for a system that will detect 95% of gunshots from rounds of .25 caliber or greater within 20 meters of the gunshot location and publish reports on them within 60 seconds. But to determine 95% detection, the agency has to define a gunshot—which leads to questions the bid doesn’t answer, Marthews said.
“The real number of unsuppressed outdoor incidents is necessarily an unknown, so how can they assess whether 95% of an unknown was detected?” Marthews said. “It is probable that what they are going to do is say a gunfire incident is what ShotSpotter declares a gunfire incident to be.”
And that reflects larger problems with ShotSpotter, according to Marthews—there are no studies showing it reduces gun crime compared to areas in which it is not used. And focusing ShotSpotter in minority areas creates a loop of detecting gunshots there and sending more police to those areas as opposed to elsewhere, he added.
“Sometimes it’s difficult for police officers to distinguish whether things advance public safety from things that advance the police department,” Marthews said.
“It creates fear in the community—BPD says there were all these gunshots last night so now I feel unsafe and want more surveillance,” Ahmad said. “Not only is it not accurate but they’re taking over the market of fear, they take all this money and it doesn’t actually reduce violence.”
The Boston City Council’s scrutiny of gunshot detection equipment is coming amidst national pushback against ShotSpotter tech. In October, officials in Dayton, Ohio, did not renew their contract with ShotSpotter because of accuracy and effectiveness concerns. In Seattle, officials decided not to contract with ShotSpotter earlier this year.
After Chicago police shot and killed 13-year-old Adam Toledo after a ShotSpotter alert in 2021, officials there found ShotSpotter does not lead to evidence of gun crimes, and activists have urged the department to stop using the company. Meanwhile, nonprofits filed a federal lawsuit saying the department has misused ShotSpotter to arrest residents, pointing to a man charged with murder after a ShotSpotter report said a gun went off near him, only for prosecutors to ultimately drop the case due to a lack of evidence.
A ShotSpotter spokesperson said, “The Chicago Police Department consistently describes ShotSpotter as a critical part of how they respond to and solve crime,” noting, “in a recent survey, 72% of Chicago residents showed support for gunshot technology.” The spokesperson added, “ShotSpotter was not responsible for Michael Williams’ [aforementioned] arrest or incarceration,” and, “Dayton PD’s announcement of their intent to not renew their contract with ShotSpotter does not negate ShotSpotter’s effectiveness. … Cherry picking a few customers that didn’t renew over the last several years doesn’t paint a realistic picture of the situation. The fact is that ShotSpotter is trusted by police departments in over 135 cities nationwide and has a 99% customer retention rate, indicating that our system works well.” You can read ShotSpotter’s full statement here.
Lara said even without the national outcry, she would be concerned about ShotSpotter and other surveillance use in Boston. She said police provided the best information about those programs they could in the time they had, but that wasn’t enough for the council to sign off on them.
“We need to have more information about the actual value this technology provides,” Lara said. “We have a responsibility that is not just fiduciary, but to make sure all services are working effectively and keeping people safe and their privacy protected.”
Crockford of ACLU Massachusetts agreed, saying councilors need more information before approving surveillance tech.
“Instead of moving forward with a committee report or vote to approve, we need to pump the brakes and slow the process down, we need a lot more hearings about this technology,” Crockford said. “Now is not the time to be committing to spending hundreds of thousands or millions of dollars on a program like this.”